with information security or privacy protection at .... A reference 2-D location, to provide the MS with ... information of the satellites to calculate the MS's position.
Mobile Positioning with AGPS In Urban Areas Shwu-Jing Chang1and Ming-Hau Ho2 1
Department of Communications and Guidance Engineering, National Taiwan Ocean University, 2, Pei-Ning Rd., Keelung, 202, Taiwan 2 China-Motor Corporation, Yang Mei, Taoyuan County, Taiwan
Abstract — Networ k-assisted GPS, or AGPS, may reduce the time-to-fir st-fix, increase the sensitivity of GPS receiver s, thus providing deeper cover age indoor s and in shadowed environments. However, with more none-line-of-sight GPS signals, quality of the position fix var ies widely. This paper systematically investigates the per for mance of var ious GPSbased positioning with simulations as well as field measurements, to fully capture the statistical char acter istics. A novel GPS-based hybr id positioning method is then proposed and evaluated with simulation. This method utilizes the AGPS assistance message to provide 100% availability with ver y promising accur acy, which makes many attr active applications feasible even in ur ban areas such as Taipei City. Key Words — GPS, DGPS, AGPS, Mobile Positioning.
I. INTRODUCTION Location has been an important asset to wireless communication network operators. With the legislative requirements on emergency mobile positioning for public safety, such as E-911 in North America and E-112 in Europe, wireless mobile positioning has gained huge research interest and resources. There are already several preliminary location-based services available to operators and users. Governments have been thinking of improving services to their people by utilizing location-based services (LBS). However, location-based services require the inter-working of several different platforms and technologies. It is quite clear that the overall cost/benefit performance or Quolity of Service (QoS) either hasn’t meet the user’s expectation or simply not attractive enough. The key factor is whether timely and accurate information can be provided to each idividual user’s need, with information security or privacy protection at appropriate cost. For geo-locating mobile users, many techniques have been proposed. Global Positioning System (GPS), Network Assisted GPS (A-GPS), Cell or cell sector identification (Cell-ID), Timing-Advance (TA), Signal Attenuation (SA) or Signal Strength (SS), Time of Arrival (TOA), Time Difference of Arrival (TDOA), Angel of Arrival (AOA), Enhanced Observed Time Difference (EOTD) are the most often mentioned methods. Usage of the location services is generally categorized in 3GPP TS 23.271 as follows: 1) Commercial LCS, or Value added Services 2) Internal LCS, for Assess Network internal operations 3) Emergency Service, provided to the emergency service provider 4) Lawful Intercept LCS, to support legally required services
Examples of possible location-based services and their horizontal accuracy requirements are listed in Table 1. Table 1 Typical LBS Required Position Accuracy Service Types Horizontal Accuracy Traffic congestion reporting 10-40m Emergency services Network based 100(67%)-300m(95%) Handset based 50(67%)-150m(95%) Person or asset tracking 10-125m Navigation 10-125m Network planning 10m –Cell ID Mobile Yellow Pages 125m-Cell ID The accuracy performances of all the proposed positioning methods are basically statistical and dependent on many factors. Comparison between positioning methods and declarations of performance based on limited field trials should be done more carefully [1-2]. Among various proposals, cell coverage based positioning methods and GPS-based positioning methods are two sets of the standard positioning methods to be supported in both UTRAN and GERAN, as specified in 3GPP TS 23.271. The organization of this paper is as follows: Section II describes various GPS-based positioning methods, including GPS, DGPS, AGPS, and the GPS positioning mechanism for mobile positioning as specified in related 3GPP/ETSI standards. Section III presents the measurement based GPS/DGPS performance analyses in open fields with unobstructed sky, and simulation-based ones in urban canyons. Design and verification of the simulation tools used in this work are also briefly described in this section. Section IV identifies information elements available for mobile positioning when timely AGPS assistance message is provided. A new AGPS-based positioning method is then proposed and analyzed. II. GPS-BASED POSITIONING METHODS
A. GPS GPS provides a means to determine position, velocity, and time around the globe. The GPS service available to civil users is called Standard Positioning Service (SPS). GPS SPS is a positioning and timing service provided on the L1 signal, containing a Coarse Acquisition (C/A) code and a navigation data message, transmitted by all GPS satellites.
The most important two official documents regarding GPS are the SPS Performance Standard [3] and the Interface Specification [4]. The Interface Specification defines the technical requirements of the interface between the GPS constellation and SPS receivers. The performance standard defines standards for SPS Signalin-Space (SIS) performance, independent of how the user applies the services provided. Uncontrollable error sources of ionosphere, troposphere, receiver, multipath, or interference are not included. GPS SPS performance parameters, i.e. service availability, service reliability, and accuracy, are statistical in nature. When the environmental and receiver error contributions are included, the global 95% horizontal error under standard conditions for single-frequency end user is estimated to be 33 meters. The service availability and accuracy can be severely degraded when the visibility of GPS satellites is blocked by the terrain and buildings, or shaded by the trees. Even with enough number of satellites in view, which is four for 3D positioning, the accuracy depends on the geometric distribution of the visible satellites. This factor is characterized as Dilution of Precision (DOP).
B. Differential GPS (DGPS) Differential GPS (DGPS) is a technique for error correction. DGPS correction data are obtained from a reference receiver, or a set of distributed reference receivers, at precisely known (surveyed) position with clear views of the sky. Errors contributed from selective availability (SA)), satellite clock, modeling of the satellite position (i.e. ephemeris), ionospheric delay, and tropospheric delay that are common to the reference receiver and the end user can be greatly reduced. SA, an intentionally introduced error source, is turned off in 2000. Currently, many coastal States are already providing standard DGPS services [5-6] to maritime users free-of-charge with DGPS broadcast near 300KHz. In order to improve the service availability and reliability, there is also a mechanism called integrity monitoring (IM) to detect unhealthy satellites and monitor the integrity of the DGPS service in both the provided DGPS corrections and the data-link for delivering such correction messages, and respond accordingly. C. Assisted GPS (AGPS) The basic idea of Assisted GPS (AGPS) is also to establish a GPS reference network (or a wide-area differential GPS network) whose receivers have clear views of the sky. This reference network is connected with the mobile communication network such as GSM for the assistance data to be transmitted/broadcast to mobile users GPS Assistance message contains three data sets: DGPS corrections, ephemeris and clock correction, almanac and other data information. Ephemeris, clock correction, almanac and other data are obtained from GPS navigation message. While the DGPS correction data in the AGPS broadcast message contains more than just the RTCM standard DGPS corrections. The following information elements are also included:
1)
A reference time, which specifies the relationship between GPS time and air-interface timing of the BTS transmission in the serving cell. 2) A reference 2-D location, to provide the MS with a priori knowledge of its location. 3) Status and health of the broadcast differential correction. If properly implemented, AGPS is expected to: 1) Reduce the time-to-first fix (TTFF) The start-up time of GPS receiver can be reduced from more than 10 minutes to within a few seconds. It is because the search space in both the code phase and frequency can be reduced, given approximate location of the handset or cell base station. 2) Increase the sensitivity of GPS receiver. As the navigation message is provided in the assistance data, longer integration time can be used to increase the sensitivity.
D. GPS Positioning Mechanism for Mobile Positioning For mobile positioning, the position calculation can be performed in the GPS-equipped target equipment (Mobile Station, MS, or User Equipment, UE) or the Mobile Location Center (MLC) in the mobile communication network. In both cases, MS/UE measures the code/phases of GPS signals. In the former case, called mobile-based, MS/UE may choose to use assistance data available from within or outside of the network during measurement and position calculation. Calculated position with an estimate of the position accuracy is then returned to the MLC. In the latter case, called network-based, the MS/UE may use available assistance data to aid the measurement process, then returns the measurements and associated quality estimates to the MLC. MLC uses these results, together with the cell-ID of the serving Base Station, position information of the satellites to calculate the MS’s position. III. GPS/DGPS PERFORMANCE ANALYSES
A. In Open Fields with Clear Views of the Sky Northern Taiwan is located in the area where 95% horizontal errors are over 10 meters for single-frequency all-in-view GPS receivers[3]. Table 2 compares the accuracy of DGPS positions with those of GPS, before and after SA was turned off. For each case in this table, 24 hours of data with one minute interval was collected with Trimble 4000 IM, a single frequency GPS receiver, at NTOU (National Tiawan Ocean University) DGPS site with clear views of the sky. This table verifies that, even without SA, GPS position accuracy is still over 10 m. Table 2 Measured GPS/DGPS Position Accuracy Date SA Horizontal Error (m) (year 2001) 68% 95% 99% RMS GPS May 1 ON 26.59 46.43 62.26 25.88 GPS May 9 OFF 5.91 10.90 13.66 3.44 DGPS Apr. 10 ON 0.2 0.29 0.35 0.13 DGPS May 11 OFF 0.18 0.25 0.29 0.10
Fig. 1 Sky plot of tracks of visible GPS satellites
Horizontal Error (m)
Tracks of visible GPS satellites as extracted from the output of a local GPS receiver indicates the lack of satellites in the northern sky of the whole Taiwan area (see Fig.1 for example). Typical measured diurnal variation (see Fig.2) indicates that the GPS positioning accuracy is degraded in the afternoon possibly by ionospheric delay, while this error source can be well removed with DGPS. Fig3 supports the arguments based on the previous two figures.
Local Time of Day (seconds) Fig. 2 Diurnal Variation in GPS/DGPS Position Errors
For the evaluation of GPS/DGPS performance in urban areas, a set of simulation tools has been developed using Matlab. This simulation tool set takes into account the characteristics of major GPS error sources. Parameters in the error statistics are set according to the local measurement data when appropriate for the simulated conditions. All the simulations presented in this work are performed with 24-hour period and one-second interval standard simulation settings. As a reference for comparison, the simulated GPS and DGPS horizontal positioning errors (95%) at NTOU are 30m and 2m, respectively. The mask angle is set at 5°. Higher receiver errors are used to simulate a low-cost GPS receiver. For DGPS simulations, ionospheric and tropospheric errors were eliminated. In cases of urban canyons, both the satellite in the line of sight (LOS) and the non-line of sight (NLOS) satellite signals with one-time reflection from the roadside buildings are included in the simulation. Simulated typical conditions in urban Taiwan are as follows: 1) North South oriented 60m-wide roads with 10 floor buildings. 2) East West oriented 60m-wide roads with 10 floor buildings. 3) North South oriented 6m or 10m wide roads with 4 floor buildings. 4) East West oriented 6m or 10m wide roads with 4 floor buildings. The simulated results are shown in Fig.4 and Table 3. It is found that DGPS can only slightly improve the horizontal position accuracy in such urban canyons. The benefit of DGPS is barely noticeable with large DOP value and the presence of none-line-of-sight signals. Position availability (with at least 3 satellites) in North South oriented 10m and 6m wide roads are only 75% and 36% respectively. The position accuracy also degrades with less available satellites.
Starlink DNAV-212G receiver 24 Hours , 1 sec. Interval
Fig.4 Simulated GPS/DGPS horizontal accuracy Table 3 Satellite Availability in 6m wide street/alley Environment Available satellites ≥4 3 2 1 0 E-W 6m wide road 5% 38% 39% 17%
